2. Addressing the Texting and Driving Epidemic:
Mortality Salience Priming Effects on Attitudes and Behavioral Intentions
Ioannis Kareklas
Assistant Professor of Marketing
Washington State University
375 Todd Addition, Pullman, WA 99164-4730
Phone: (509) 335-2781
Fax: (509) 335-3865
ioannis.kareklas@wsu.edu
Darrel D. Muehling
Chair and Professor of Marketing
Washington State University
367C Todd Addition, Pullman, WA 99164-4730
Phone: (509) 335-7302
Fax: (509) 335-3865
darrel@wsu.edu
Note: Authorsâ names are listed alphabetically. Each contributed equally to this manuscript.
3. F
o
r
R
e
v
i
e
w
O
n
l
y
Addressing the Texting and Driving Epidemic:
Mortality Salience Priming Effects on Attitudes and Behavioral Intentions
ABSTRACT
Texting while driving is becoming a problem of epidemic proportion, causing thousands of
fatalities each year. However, surprisingly few academic studies to date have examined this issue
in a social marketing context. We address this research void by reporting the findings of two
empirical studies â the first, an exploratory study of driversâ perceptions of texting while driving;
the second, an experimental examination of the relative effectiveness of mortality salience
primes in public service announcements (PSAs). Employing theory derived from the mortality
salience literature, we find that when verbal and/or visual cues to death/dying were used,
participantsâ attitudes and behavioral intentions were altered in a positive direction. As compared
to a control group, the primed PSAs produced less favorable attitudes and reduced intentions to
text while driving in the future. Implications of these findings for consumers, social marketers
and policy makers are discussed, and future research directions are provided.
Page 1 of 59 Journal of Consumer Affairs
4. F
o
r
R
e
v
i
e
w
O
n
l
y
1
Distracted driving â defined as âany activity that may divert a personâs attention away
from the primary task of drivingâ (NHTSA 2013) â has recently become a matter of great
concern. Of all the types of driver distractions (e.g., talking to passengers, grooming, adjusting a
radio, etc.), text8messaging is often considered the most distracting â and potentially dangerous â
because it requires visual, manual, and cognitive attention from the driver (Dingus, Hanowski,
and Klauer 2011; Drews et al. 2009). In support of this notion, the National Safety Council
(2010) estimates that at least 1.6 million crashes each year (about 28% of all traffic accidents)
and thousands of fatalities can be attributed to drivers who are distracted by cell phone use, with
conservative estimates indicating that as many as 200,000 of these accidents are due specifically
to texting8while8driving incidents.
Several approaches have been recommended and initiated over the past few years in an
attempt to remedy what some have suggested has become an âepidemicâ problem in the United
States (OSHA 2010). Among these are: (1) federal and state laws banning the practice, (2)
stricter enforcement of current laws, (3) company policies restricting or prohibiting employee
cell phone usage while driving, (4) advanced technologies (e.g., preventing incoming texts and
calls from reaching a driverâs cell phone), (5) the use of a âdesignated (passenger) texter,â and,
(6) the dissemination of public service announcements (PSAs) and other promotional campaigns
aiming to dissuade drivers from engaging in these risky behaviors.
Although regulatory, company8mandated, and technological approaches would appear to
be promising avenues for addressing this growing concern, government findings suggest that the
incidence of distracted driving accidents has not diminished in a significant way. In fact, research
conducted by the Highway Data Loss Institute (2010) suggests that enacting texting bans may
actually increase the likelihood of an accident occurring (because, to avoid being caught, drivers
Page 2 of 59
Journal of Consumer Affairs
5. F
o
r
R
e
v
i
e
w
O
n
l
y
2
often move their phones down and out of sight when they text â further exacerbating the inherent
risks). Such findings support the notion that continued research efforts on the topic are sorely
needed. As such, the current research was undertaken with a specific focus on providing
additional insights regarding the use of persuasive communication techniques that may
successfully alter perceptions and curb the incidence of texting while driving. Recognizing that
media8delivered promotional campaigns continue to be viable vehicles for educational and
persuasive purposes, we examine the relative effectiveness of targeted fear8based advertising
appeals in the specific context of public service announcements. Mortality salience (Greenberg,
et al. 1990) is offered as a theoretical basis for our investigations.
We begin with a review of the current state of the texting8while8driving problem, which
includes a discussion of the physiological impact of texting8while8driving, and the psychological
factors that may contribute to this behavior. Then, we provide the methodological details and
discuss the findings of the first of two empirical studies that were undertaken. In Study 1 (an
exploratory study of driversâ perceptions of texting while driving), we identify some common
factors that contribute to driversâ attitudes, intentions, and texting8while8driving behaviors â
factors we believe to be promising avenues to pursue further in a social marketing context.
Following the insights gleaned from this preliminary study, we then provide a review of the fear
appeals literature (focusing on mortality salience), which guides our research exploration in the
second study. Study 2 investigates the differential effects that a verbal (written statement) and/or
visual (skull8and8crossbones) mortality salience prime may have on recipientsâ attitudes and
intentions to text and drive. Implications for social marketers are then discussed, along with
several directions worthy of future research pursuit.
Page 3 of 59 Journal of Consumer Affairs
6. F
o
r
R
e
v
i
e
w
O
n
l
y
3
BACKGROUND AND LITERATURE REVIEW
Whether watching television, reading the newspaper, or surfing the web, it is quite
apparent that texting while driving has become an issue garnering considerable national attention
in recent years. As examples, the month of April is now recognized as âNational Distracted
Driving Awareness Monthâ by the National Highway Traffic Safety Administration (NHTSA)
and the National Safety Council (NSC), and public service announcements using the tagline
âOne text or call could wreck it allâ have been developed in an attempt to raise awareness of and
help curb cell phone usage. The impact of distracted driversâ behaviors is also a central theme
played out in Ad Council/NHTSA videos appearing on the Internet (NHTSA 2013). Recently,
during the 2012 Summer Olympics, several iterations of AT&Tâs âIt Can Waitâ campaign were
aired, highlighting the dangers of texting while driving. At the same time, NBC aired the pilot
episode for a new program (âGo Onâ) starring Matthew Perry. Although this program is intended
to be comedic in nature, its premise is that Matthew Perryâs character is seeking grief counseling
due to his wifeâs sudden death in a texting8while8driving car accident.
In addition, several non8profit organizations (e.g., the American Automobile Association
(AAA), the National Safety Council, the Ad Council, Stop Texting and Driving (S.T.A.D.)) and
insurance companies (e.g., Allstate, Liberty Mutual, Nationwide, and State Farm) have created
websites and/or sponsored advertising campaigns warning drivers of the dangers of texting, as
well as asking members, employees and customers to sign pledges to commit to distraction8free
driving. This issue has become so alarming that 41 U.S. states to date have passed legislation
banning text8messaging8while8driving for all drivers, and several additional states have signed
partial bans and restrictions (NHTSA 2013).
Page 4 of 59
Journal of Consumer Affairs
7. F
o
r
R
e
v
i
e
w
O
n
l
y
4
A number of academic studies and government8sponsored reports have documented the
dangers related to texting while driving. Regarding the physiological impact, reports suggest that
using a cell phone while driving delays driversâ reactions as much as having a blood alcohol
level at or above the legal limit of .08% (Strayer, Drews, and Crouch 2006). Furthermore, text8
messaging has been shown to be associated with the highest level of potential distraction
(Dingus, Hanowski, and Klauer 2011; Ranney et al. 2012). For example, studies show that
drivers who retrieve or send text messages are four times more likely to be in an accident serious
enough to injure themselves and/or others (Hosking, Young, and Regan 2006; Redelmeier and
Tibshirani 1997). Additionally, using functional magnetic resonance imaging (fMRI),
researchers at Carnegie Mellon University found that brain activity related to driving decreases
by 37% when using a cell phone (Just, Keller, and Cynkar 2008).
Despite these reported dangers, research indicates that as many as 75% or more of drivers
report engaging in distracted driving behaviors (Atchley, Atwood, and Boulton 2011; Nelson,
Atchley, and Little 2009). A recent study by Consumer Reports (2012) supports this finding in
the context of texting while driving; while 80% of young drivers agreed that texting while
driving is dangerous, about a third of them admitted to doing so in the past month. Furthermore,
Lee et al. (2008) found that younger drivers are likely to continue text8messaging even when
they are faced with challenging driving situations. Consistent with commonly held beliefs,
respondents under the age of 25 reported a higher likelihood of sending text messages while
driving than did older drivers. Additionally, according to this study, the primary reason for
sending a text message (across gender and age groups) was the level of importance of the
message.
Page 5 of 59 Journal of Consumer Affairs
8. F
o
r
R
e
v
i
e
w
O
n
l
y
5
Research has also begun investigating the psychological factors contributing to distracted
driving behaviors, including texting while driving. This emerging research stream has focused on
individualsâ perceptions of riskiness (Atchley, Atwood, and Boulton 2011), insufficient sleep
(Dahl 2008), recklessness and its consequences (Harrison 2011), compulsiveness (Steelman et al.
2012), and addiction (Bianchi and Philips 2005; Takao, Takahashi, and Kitamura 2009). As an
example, a sampling of college students by Harrison (2011) found that most believed texting
while driving is distracting and dangerous, and that it should be illegal. Ironically, the majority of
the same participants admitted to texting while driving at least occasionally â most of whom
admitted to doing so even while passengers were present in their vehicles. Furthermore,
participantsâ attitudes toward texting while driving were significantly correlated with other
reckless driving behaviors (e.g., veering into another driving lane, speeding, etc.).
Other research on the topic suggests that perception of risk may be a weak predictor of
reading, replying to, or initiating texts while driving among young drivers (Atchley, Atwood, and
Boulton 2011). Findings such as these have led some scholars to suggest that young drivers tend
to believe that texting8related traffic accidents (and even death) may occur to others, but will not
happen to them (Magid 2009). Nonetheless, employing the Theory of Planned Behavior (Azjen
1991), Nemme and White (2010) found that both behavioral attitudes and subjective norms were
predictive of young driversâ intentions to text while driving. Group norms (i.e., perceptions of
their reference groupâs texting8while8driving behaviors), and moral norms (i.e., the perceived
moral correctness/incorrectness of the behavior) added explanatory power to their model.
Most relevant to the current investigation, three studies to date have examined fear8based
emotional appeals and their impact on various types of driving behaviors. Lewis, Watson, and
Whiteâs (2010) study confirmed the relevance of cognitive and emotional components of
Page 6 of 59
Journal of Consumer Affairs
9. F
o
r
R
e
v
i
e
w
O
n
l
y
6
persuasive messages in the context of speeding behaviors. More specifically, ad8evoked fear and
anxiousness were shown to have an impact on both acceptance and rejection of anti8speeding
messages. Furthermore, participants were more likely to be persuaded by messages when they
believed they possessed the ability to enact a useful strategy that would reduce the posed threat
of being involved in a speeding8related car accident.
Focusing on four unsafe driving behaviors (including texting while driving), Lennon,
Rentfro, and OâLeary (2010) observed an unintended effect in their study of fear appeals directed
at young drivers. Specifically, after viewing fear8based PSAs designed to discourage distracted
driving behaviors, participants reported a likelihood of engaging in these unsafe
practices. The authors suggested that the low/moderate level of fear used in their experiment was
perhaps not great enough to produce the anticipated results.
As a follow8up to this research, Lennon and Rentfro (2010) conducted focus groups of
young adults, and recorded their responses to six PSAs employing fear appeals (one of which
focused on distracted driving). Their findings indicated that fear (graphic content and fear
arousal), perceived threat (the perceived likelihood and severity of consequences), and perceived
efficacy were among the most influential predictors of the effectiveness of the fear8based
appeals.
From our overview of the relevant literature, it is apparent that while anti8texting8while8
driving promotional campaigns are quite prevalent in the media today, few if any academic
studies to date have sought to examine empirically their potential effectiveness in changing
driversâ attitudes and intentions in a normative direction advocated by social marketers. The
studies we report below were conducted in an attempt to help address this research void and to
identify effective means of deterring young drivers from texting while driving.
Page 7 of 59 Journal of Consumer Affairs
10. F
o
r
R
e
v
i
e
w
O
n
l
y
7
STUDY 1
We first conducted an exploratory study to obtain base8line attitudinal and behavioral
intention measures and to further examine the reasons why many young drivers text and drive,
while others do not. It was our intention to corroborate observations reported in the media to date
and to identify factors that could be potentially manipulated and further examined in the context
of public service announcements.
A nationally representative sample of drivers between the ages of 18 and 49 was obtained
from an on8line panel provider ( = 357). The average age of respondents was 24.9 years old,
similar to the age of participants from other texting8while8driving research (e.g., Atchley,
Atwood, and Boulton 2011; Hosking, Young, and Regan 2006; Nemme and White 2010), and
consistent with a range of ages making up the prime target demographic for social marketers in
the current context. Demographic measures indicated our sample was 51.0% female and 74.8%
Caucasian/White (10.4% Black/African8American, 9.0% Asian8American, 4.2% Hispanic, 1.6%
other races/ethnicities). In terms of geographic location, 31.1% of participants reported living in
the South, 25.5% in the Northeast, 21.8% in the Midwest, and 21.6% in the West â providing a
reasonable cross8section of the nation.
The first portion of the survey consisted of questions focusing on attitudes and intentions,
adapted from standard measures reported in the marketing literature (cf. Spears and Singh 2004).
Specifically, participantsâ attitudes were assessed by asking âWhat is your attitude toward
texting while driving?,â followed by four 78point bipolar adjective pairs of items (âBad/Good,â
âFoolish/Wise,â âHarmful/Beneficial,â âNot acceptable/Acceptableâ). An
scale (AB) was subsequently created from these responses (Cronbachâs α = .94), with
Page 8 of 59
Journal of Consumer Affairs
11. F
o
r
R
e
v
i
e
w
O
n
l
y
8
larger values indicating more favorable attitudes toward texting while driving. Similarly,
respondentsâ intentions to text and drive in the future were assessed by asking âWhat is the
likelihood that you will text while driving over the next month?,â followed by two 78point
bipolar adjective pairs (âNot at all likelyâ (1)/âVery likelyâ (7) and âExtremely unlikelyâ
(1)/âExtremely likelyâ (7)). These items were combined to create a (BI)
scale (α = .98), with larger values indicating a greater propensity to text while driving.
Our subsequent analysis of responses to these items indicated that participants held
relatively negative attitudes toward texting while driving ( AB = 1.77), and reported relatively
low propensities to text while driving in the future ( BI = 2.59). Nonetheless, a significant
minority (24.9% of individuals surveyed) indicated that they planned to text and drive over the
next month. In addition, we found that participantsâ attitudes and behavioral intentions were
significantly correlated ( = .58, < .001), consistent with attitudinal research which suggests
attitudes are often predictive of behavioral intentions (Ajzen 1991; Ajzen and Fishbein 1980).
As a follow up to the behavioral intention questions, participants were also asked the
extent to which they would maintain their current pattern of behavior, by responding to the
statement: âI do not plan to change my texting8while8driving behavior in the near future,â with
scale endpoints âStrongly disagreeâ (1)/âStrongly agreeâ (7). This item was accompanied by an
open8ended question, asking respondents to explain why they answered in the way they did.
From these measures, we were able to compare the responses of those individuals who were
likely to continue to text and drive from those who were likely to continue to refrain from doing
so in the future.
Page 9 of 59 Journal of Consumer Affairs
12. F
o
r
R
e
v
i
e
w
O
n
l
y
9
A common theme emerged from our examination of participantsâ responses. The most
prevalent reason why respondents indicated they did not intend to text while driving in the future
is that they found this behavior to be very . Typical comments included:
âKids my age who text and drive are not paying attention to the road and cause
accidents. I lost my friend to this.â
âI know it is dangerous and could cause a serious or even fatal accident. It is
unsafe for a driver to text while driving. It is worse than drunk driving.â
âI believe it to be an impairment of oneâs ability to drive equal to or greater than
that of alcohol.â
âI believe texting while driving is extremely dangerous. It only takes a second for
an accident to happen.â
âI think it is very distracting, making it extremely unsafe.â
âI am aware of the dangers, and a text is not worth my life.â
Interestingly, participants who indicated that they were likely to continue to text while
driving in the future also made reference to danger. However, while this group acknowledged
their behavior was potentially dangerous, they rationalized that they were able to maintain
control of their cars while text8messaging. Specifically, several respondents indicated that they
believed they were skilled enough to be able to safely text and drive at the same time. Sample
comments included:
âI know itâs a terribly dangerous thing to do. ButâŠI always justify within myself
that Iâm just texting a few words, itâs no big deal. I can handle this.â
âI only glance long enough to read a word or two, look at the road, glance again,
and so on. This isnât that dangerous.â
âI use one hand to text and one hand to drive so I maintain control of the car.â
âI think I can text and drive without any negative consequences.â
âI am just so used to it and I think I wonât have a wreck.â
âI believe that I am a driver who is aware enough of his surroundings to be able to
text while driving without causing any incidents or accidents.â
Page 10 of 59
Journal of Consumer Affairs
13. F
o
r
R
e
v
i
e
w
O
n
l
y
10
We also asked participants to provide their evaluative judgments regarding several
hypothetical situations related to texting8while8driving behaviors (all 78point items with scale
endpoints: âStrongly disagreeâ (1)/âStrongly agreeâ (7)). For illustrative purposes, these results
are separately reported in Table 1 for participants who indicated that they planned to text while
driving over the next month ( = 174) and those who indicated that they did not plan to text
while driving over the next month ( = 183). This was accomplished by using a median split
(i.e., a relative measure) of participantsâ likelihood to text and drive in the future by separating
responses above and below the median of 1.5 on the BI scale. Additionally, this table reports the
correlation coefficients between responses to each of the hypothetical situations and the
scale in its continuous form. Comparable analyses are also reported using
the scale.
[Insert Table 1 about here]
The mean responses of participants who indicated that they planned to text and drive in
the near future (as compared to those who indicated that they did not plan to text and drive) were
significantly different (all 8values < .001). As compared to the latter group, the former group
believed that it was more acceptable to text and drive under a variety of situations such as: (a)
when stopped at a red light, (b) in case of an emergency, (c) when there is little traffic on the
road, (d) to get/give directions, (e) when they wanted to know something instantly, (f) when they
thought they can get away with it, (g) when responding to messages from their friends and
family, and (h) when they are bored. Nonetheless, both participant groups tended to agree with
the statement that if they texted while driving they would: (a) kill themselves, and (b) kill others
in a crash (each mean value significantly greater than 4.0 on the 78point scale, < .001) â thus
offering some initial evidence in support of the use of mortality salience (i.e., making individuals
Page 11 of 59 Journal of Consumer Affairs
14. F
o
r
R
e
v
i
e
w
O
n
l
y
11
aware of their eventual death) as a persuasive priming cue. However, it should be noted that the
mean responses to both statements were statistically different between the two respondent groups
(both 8values < .001), indicating that those individuals who planned to text and drive in the
future (as compared to those who did not plan to do so) perceived a lower likelihood of killing
themselves and/or others in a crash if they were to text and drive.
Participantsâ responses to each of the hypothetical situations were also significantly
related to their responses to the and scales
(all correlation coefficients were significant at < .001; see Table 1). To further investigate the
nature of the relationships between these constructs, we conducted a mediation analysis
(ordinary least squares path analysis) using bootstrapping to test the indirect effects (Preacher
and Hayes 2004; 2008; Zhao, Lynch, and Chen 2010). was used as
the independent variable, as the dependent variable, and a summary index
of the (EJ) as the mediating variable (see Figure 1). The EJ index was
created by combining participantsâ responses to each of the 13 hypothetical situations (α = .95,
after reverse coding the 4 items shown in the bottom panel of Table 1). Results indicated that
participantsâ attitudes significantly influenced their evaluative judgments, and these judgments
(while controlling for AB) predicted their behavioral intentions. A bias8corrected bootstrap
confidence interval for the indirect effect of AB on BI (ÎČ = .53) using 1,000 bootstrap samples
was completely above zero (.45 to .65). There was no evidence that AB influenced BI,
independent of its indirect effect through EJ, as the direct effect of AB on BI was not statistically
significant (ÎČ = .05; = .26).
[Insert Figure 1 about here]
Page 12 of 59
Journal of Consumer Affairs
15. F
o
r
R
e
v
i
e
w
O
n
l
y
12
THE FOCUS OF OUR SUBSEQUENT INVESTIGATION
In addition to helping to delineate the relationship between attitudes, evaluative
judgments, and behavioral intentions, the findings of Study 1 offer evidence that perceptions of
imminent danger vary across groups of individuals who intend to text while driving in the future
from those who do not. Additionally, we observed that dangerous outcomes are often associated
with injury or death. Therefore, in our next study, we operationalized and examined this focus on
dangerous outcomes in the context of death/dying. To do so, we drew upon the literature related
to mortality salience. Our assumptions (tested empirically in Study 2) were that if proper
mortality salience primes (i.e., ad cues) were employed, thoughts of death could be evoked, and
desirable attitudinal and behavioral intention outcomes (i.e., more negative attitudes toward the
behavior and lower intentions to text and drive in the future) could be attained.
Marketers have commonly used one of two types of appeals to discourage socially
undesirable behaviors: ârational/informationalâ and âemotional/fearâ appeals (Lennon, Rentfro,
and OâLeary 2010). Rational appeals are typically characterized as informational in nature â
informing individuals by presenting logical message arguments that advocate a certain position.
On the other hand, emotional appeals attempt to enhance the persuasiveness of a message by
eliciting an emotional reaction (such as anger, sadness, or fear) that is intended to alter existing
predispositions (Perse, Nathanson, and McLeod 1996).
In the current context, the inherent risks associated with texting while driving could be
highlighted by using either a rational or an emotional appeal. However, some researchers have
suggested that emotional appeals may be more effective than rational appeals in dealing with a
variety of social issues (Terblanche8Smit and Terblanche 2010). Empirical findings have
Page 13 of 59 Journal of Consumer Affairs
16. F
o
r
R
e
v
i
e
w
O
n
l
y
13
supported this assertion in the context of health campaigns designed to address AIDS/HIV (Flora
and Maibach 1990) and more closely related to our investigation, road safety (Elliott 1993). This
research stream indicates that emotional appeals may more effectively engage recipients on a
personal level, attract more attention, and are generally more memorable (Flora and Maibach
1990; Petty and Cacioppo 1986). In addition, research has shown that individuals who are
skeptical of advertising (due, in part, to their negative perceptions of advertisersâ manipulative
intent; Tien and Phau 2010) tend to respond more positively to emotional appeals than to
informational appeals (Obermiller, Spangenberg, and MacLachlan 2005).
Therefore, in our next study, we examine whether the use of emotional appeals that relate
the dangers of texting while driving to death/dying may be effectual in altering recipientsâ
attitudes and behavioral intentions toward texting while driving. We begin our discussion of
mortality salience by providing a brief overview of the fear appeals literature in the context of
health communications. Then, using the Extended Parallel Process Model (Witte 1992) as a
conceptual base, we discuss the key components and processes that contribute to individualsâ
responses to fear appeals specifically related to texting while driving. This discussion concludes
with a review of the literature pertaining to reactance and manipulative intent, which suggests
that the impact of social marketing communications may be potentially undermined if recipients
feel defensive and object to the ad sponsorâs use of fear appeals.
CONCEPTUAL DEVELOPMENT
Fear Appeals
âFear is an emotional response to a threat that expresses, or at least implies, some sort of
dangerâ (Tanner, Hunt, and Eppright 1991, 36). In keeping with this notion, social psychology
Page 14 of 59
Journal of Consumer Affairs
17. F
o
r
R
e
v
i
e
w
O
n
l
y
14
researchers have developed several theories and models designed to explain, predict, and explore
the impact of fear on consumersâ attitudes, intentions, and behaviors. These include drive
theories (Hovland, Janis, and Kelly 1953), parallel process models (Leventhal 1970), subjective
expected utility (SEU) models such as the protection motivation theory (PMT; Rogers 1975), as
well as extended parallel process models (EPPM; Witte 1992). The integration of these theories
reveals four key components of successfully executed fear appeals: (a) the evocation of fearful
emotions (typically related to the perceived severity of the threat); (b) the perceived probability
of occurrence of the threat; (c) perceived response efficacy (i.e., individualsâ perceptions
regarding whether the recommended response would be successful if implemented); and (d)
perceived self8efficacy (i.e., individualsâ belief in their ability to take preventative action) (de
Hoog, Stroebe, and de Wit 2007; Rogers 1983).
To guide the present discussion, Figure 2 depicts the interrelationships between these key
constructs and their resulting outcomes in the context of texting while driving. This conceptual
model is an adaptation of Witteâs (1992) Extended Parallel Process Model (EPPM). We chose to
focus on the EPPM because, as discussed by Witte and Allen (2000), this model integrates
previous theoretical perspectives in the fear appeals literature into one unified theory.
[Insert Figure 2 about here]
Extant research indicates that persuasive messages that focus on fear, especially the fear
of oneâs own death, may be especially effective at altering individualsâ attitudes and behaviors
(Greenberg et al. 1990; Keller and Lehman 2008; Witte and Allen 2000). Among these, several
recent studies have investigated how fear appeals may be useful in addressing social issues such
as smoking (Kees et al. 2006; 2010; Reardon et al. 2006), HIV/AIDS (Dahl, Frankenberger, and
Manchanda 2003), alcohol abuse (Shehryar and Hunt 2005), and drug use (Passyn and Sujan
Page 15 of 59 Journal of Consumer Affairs
18. F
o
r
R
e
v
i
e
w
O
n
l
y
15
2006), among others. As an example, Kees et al. (2010) investigated the effectiveness of graphic
pictorial warnings placed on cigarette packaging as a smoking deterrent. These researchers found
that pictorial warnings were effective due to their ability to evoke fear, which mediated the effect
on intentions to quit smoking. Relatedly, Keller and Lehmanâs (2008) meta8analysis on the
effective use of fear in health communications found messages that stress social and physical
consequences have the best chance of increasing pro8health intentions.
We focus on emotional fear8based appeals for several reasons. Teenagers and young
adults often believe they are invincible, and thus, tend to engage in dangerous behaviors without
considering the inherent risks (Wickman, Anderson, and Greenberg 2008). Furthermore, a recent
meta8analysis reported that there has been a systematic increase in narcissism among college8
aged individuals in recent years (Twenge et al. 2008), which is often associated with risky
decision making (Campbell, Goodie, and Foster 2004). Such findings suggest that some drivers â
especially younger drivers â may believe their driving abilities are not impaired by distractions
such as texting while driving, and that they may continue to engage in this very dangerous
behavior, even though they recognize its associated risks (which they rationalize away). In fact,
the findings of our exploratory study, in addition to several related reports (e.g., Tison,
Chaudhary, and Cosgrove 2011) support this very notion. Specifically, Study 1 respondents who
admitted to texting and driving in the past, also believed that it was unlikely that their behavior
would lead to an accident in the future, as they believed in their ability to text and drive safely.
This mindset is consistent with past reports indicating that college8aged individuals tend to
exhibit distorted judgments of their own abilities, leading them to engage in dangerous behaviors
(Paulhus et al. 2003). Misperceptions of their actual driving skills and associated increases in
Page 16 of 59
Journal of Consumer Affairs
19. F
o
r
R
e
v
i
e
w
O
n
l
y
16
narcissism and perceived invincibility in the past few years appear to be contributing factors in
young driversâ decision to text and drive.
One method through which the impact of fear8based emotional appeals is strengthened is
by reminding individuals of their own mortality. Social psychological theories suggest that
nearly all human actions are motivated (either directly or indirectly) by the awareness of oneâs
own mortality (Greenberg et al. 1990). According to Terror Management Theory (Greenberg,
Pyszczynski, and Solomon 1986), humans are conflicted because they possess the instinct to
avoid death, but also have the intellectual capacity to recognize that attempts to avoid death will
ultimately be futile. Indeed, extant research has demonstrated that drawing attention to oneâs
mortality prompts thoughts of death and dying and can have a profound impact on individualsâ
attitudes and behaviors. For example, a recent meta8analysis of the mortality salience (MS)
literature revealed that priming people with the idea of death produces robust, âmoderate to large
effects across a wide variety of MS manipulations as well as attitudinal, behavioral, and
cognitive DVsâ (Burke, Martens, and Faucher 2010, 187).
In the current context, we argue that following exposure to a mortality salience prime, the
perceived severity of the threat posed by texting while driving will increase, as it is associated
with the individualâs underlying fear of death (Keller and Lehman 2008; Routledge and Juhl
2010). This perception should in turn increase oneâs motivation to develop a means to âmanage,â
and thereby, reduce the perceived threat (Pavia and Mason 2004). Mortality salience consciously
or subconsciously motivates individuals to take actions to avoid threatening behaviors that may
cause them to die, or to distract themselves from thinking about their own mortality. In the
Page 17 of 59 Journal of Consumer Affairs
20. F
o
r
R
e
v
i
e
w
O
n
l
y
17
context of the current investigation, one means of alleviating the threat is by choosing to not text
and drive (consistent with decreased attitudes and intentions to engage in this dangerous
behavior in the future). Perceptions of the severity of the threat (assuming a PSA effectively
evokes thoughts of mortality) is also expected to help mitigate the tendency of individuals to
discount persuasive messages. The fear appeals literature, and in particular the extended parallel
process model (Witte 1992) provides theoretical support for these suppositions.
According to the EPPM, recipients first appraise the severity of the threat. If perceived
threat is high (as it should be in the case of death/dying), the level of perceived efficacy (self8
efficacy and response efficacy) will determine whether individuals use âfear controlâ or âdanger
controlâ strategies to deal with the perceived threat (Witte, Meyer, and Martell 2001). If both
perceived threat and efficacy are high, individuals are expected to focus on ,
whereby they attempt to deal with the threat by contemplating possible solutions that will allow
them to reduce/eliminate the threat (i.e., decide to avoid texting and driving). If instead,
perceived threat is high, but self8efficacy and/or response efficacy are low, then individuals may
instead focus on , which involves the use of âmaladaptive coping mechanismsâ to
reduce fear (Witte 1992; Witte and Allen 2000). Such coping mechanisms include denial and
rationalization (similar to our Study 1 respondents who argued that they could text and drive
safely), which involve discounting the advertiserâs message and thereby having no associated
change in attitudes and intentions, or psychological reactance (discussed in the next section)
(Witte 1992; Witte and Allen 2000; Witte et al. 2001).
We argue that both self8efficacy and response efficacy should typically be high in the
case of texting while driving, as drivers can choose to not engage in this dangerous behavior
(self8efficacy), and doing so should effectively reduce the associated threat of death (response
Page 18 of 59
Journal of Consumer Affairs
21. F
o
r
R
e
v
i
e
w
O
n
l
y
18
efficacy). Therefore, we expect that exposure to a mortality salience prime within a PSA,
coupled with the concomitant realization that individuals have control over their own behaviors,
will lead to less favorable attitudes and reduced intentions to text while driving in the future.
However, this assumes that individuals will accept the advertiserâs position as being balanced
and fair, and not manipulative in nature. If individuals perceive the ad message to be unfair or
biased, or believe the ad sponsor is attempting to unduly persuade them via manipulative means,
they may instead react adversely, which could hinder message acceptance and persuasiveness â a
point we elaborate upon in the following section.
When readers/viewers believe that an advocated position is threatening or eliminating
their own behavioral freedoms, some level of (Brehm 1966) is expected.
Reactance theory suggests that when an individual feels pressured to accept a certain point of
view, s/he is likely to respond by adopting or strengthening a mindset that is to what is
intended (Brehm 1966; Brehm and Brehm 1981). In advertising research, an individualâs
perception that a marketer/advertiser is attempting to persuade them (perhaps in an unfair
manner) is often referred to as (Campbell 1995).
Similar to reactance theory, recent findings from this body of research suggest that when
perceived manipulative intent is high, the persuasive effectiveness of the communication may be
undermined (Cotte, Coulter, and Moore 2005). If psychological reactance is high, a âboomerangâ
outcome might occur, whereby recipients exhibit an unexpected greater (rather than lesser)
propensity to practice unsafe driving. This general pattern of effects is consistent with a
curvilinear (i.e., an inverted U8shaped) relationship between arousal of fear and persuasion
Page 19 of 59 Journal of Consumer Affairs
22. F
o
r
R
e
v
i
e
w
O
n
l
y
19
(Keller 1999; Krisher, Darley, and Darley 1973; Quinn, Meenaghan, and Brannick 1992;
Sternthal and Craig 1974), where a high level of fear might lead to âmessage derogationâ and/or
a âdenial of the threatâ (Ruiter, Abraham, and Kok 2001). Therefore, for PSAs featuring
mortality salience primes to be most effective, they must not unduly prompt perceptions of
unfairness or manipulative intent.
STUDY 2
The purpose of Study 2 was to determine whether and to what extent visual cues (i.e., the
insertion of a skull8and8crossbones symbol) and/or verbal cues (i.e., a written statement
referencing death) in PSAs might be effective in changing respondentsâ attitudes and intentions
to text while driving in the future. To provide additional insights for social marketers
endeavoring to address the texting8while8driving epidemic, we also explored whether
âmanipulative intentâ and participantsâ opinions of the PSAs influenced our observed results.
Method
A nationally representative sample of individuals between the ages of 18 and 49 was
again obtained from an on8line panel provider ( = 224). The average age of respondents was
24.3 years old. The sample was 44.2% female, 72.3% Caucasian/White (11.2% Asian8American,
8.5% Hispanic, 4.5% Black/African8American, and 3.5% Other), with 27.4% living in the South,
23.8% in the Northeast, 21.0% in the Midwest, and 27.8% living in the West. Similar to Study 1,
we obtained a reasonable cross8section of the nation, with respondentsâ demographic
Page 20 of 59
Journal of Consumer Affairs
23. F
o
r
R
e
v
i
e
w
O
n
l
y
20
characteristics comparable to individuals who are typically targeted for anti8texting8while8
driving campaigns. Participants were randomly assigned to one of four treatment conditions in a
2 (Verbal Cues: Present vs. Absent) x 2 (Visual Cues: Present vs. Absent) between8subjects
design.
To test our underlying theoretical assumptions, four public service announcements were
developed. All PSAs were presented at a resolution of approximately 1,024 x 768 pixels and
contained the same image (a full8color picture of a driver texting while driving), the headline
âTexting While Driving: A Dangerous Combination,â and brief ad copy suggesting âPlease donât
text and drive.â The PSA used in the control (verbal/visual absent) condition was designed to act
as a baseline for measuring the incremental impact of adding mortality salience primes (either a
verbal, a visual, or both a verbal and a visual reference to death in the context of texting while
driving). In the verbal condition, the following ad copy emphasizing the deadly nature of texting
and driving was inserted: âTexting while driving kills. The National Highway Traffic Safety
Administration estimates that texting while driving kills 3,000 people every year.â In the visual
condition, a skull8and8crossbones symbol (selected following a pretest, as described below) was
prominently displayed in the center of the PSA. The final PSA condition incorporated the skull8
and8crossbones symbol, in addition to the ad copy noted above (see Appendix).
In preparation for Study 2, we conducted two pretests to determine an appropriate visual
death symbol to prime participantsâ thoughts of mortality. This was done in two separate phases.
Page 21 of 59 Journal of Consumer Affairs
24. F
o
r
R
e
v
i
e
w
O
n
l
y
21
In phase 1, sixty8five undergraduate students were asked to complete a free8elicitation task by
listing at least five objects or symbols that personally reminded them of death. Among the
symbols they listed (crosses, coffins, tombstones, ghosts, the grim reaper, etc.), the skull8and8
crossbones symbol dominated, and therefore was chosen for further testing in the second phase
of stimulus development.
In phase 2, fifty8two student participants were asked to assess the extent to which a skull8
and8crossbones symbol shown to them evoked thoughts of death. Seven items were used to make
this assessment. Two Likert8scaled (âStrongly disagreeâ (1)/âStrongly agreeâ (7) statements:
âThis symbol reminds me of deathâ and âThis symbol represents deathâ) were used, as well as
five 78point bipolar8adjective items (âThis symbol is more representative ofâŠâ âAlive/Dead,â
âBeginning/End,â âLiving/Dying,â âGo/Stop,â and âBirth/Deathâ). Based upon the internal
consistency of the items (α = .91), an index was subsequently created. Analysis of participantsâ
responses ( = 5.69, well above the 4.0 scale midpoint) justified use of the skull8and8crossbones
symbol as a visual mortality salience prime.
The literature on the effectiveness of signs has shown that familiar symbols are processed
more quickly than written text (Taylor, Claus, and Claus 2005; U.S. Small Business
Administration 2003). For example, research shows that individuals require about 0.33 seconds
to recognize a familiar symbol or a single word, whereas written statements containing multiple
words take much longer to process (U.S. Small Business Administration 2003). Focusing on
anti8texting8while8driving PSAs, our experimental stimuli allowed us to further examine whether
symbols (i.e., skull8and8crossbones) may more effectively evoke thoughts of death (and thereby,
influence attitudes and behavioral intentions) â more so than written messages that similarly
make references to death and dying.
Page 22 of 59
Journal of Consumer Affairs
25. F
o
r
R
e
v
i
e
w
O
n
l
y
22
Participants were randomly assigned to one of the four PSA conditions and were allowed
30 seconds to view it before responding to a series of attitudinal and behavioral intention
measures. The same measures of attitudes (AB) (α = .93) and behavioral intentions (BI) (α = .97)
toward texting while driving, reported in Study 1, were once again employed (with smaller
values indicating outcomes in line with social marketersâ goals). In addition, participants were
asked to respond to manipulation8check items corresponding to their perceptions of the PSAâs
effectiveness in evoking thoughts of death. These items included: âTo what extent did the PSA
make you think of death?â (âNot at allâ (1)/âA lotâ (7)), as well as âPlease indicate your level of
agreement with the following statements:â âThis PSA reminds me of death,â and âThis PSA
invokes thoughts of death,â with endpoints âStrongly disagreeâ (1) and âStrongly agreeâ (7). The
three items were subsequently combined to form a mortality salience index (α = .94).
Additional measures of participantsâ responses to the PSAs were also taken to help us
gain further insights regarding the pattern of findings observed for AB and BI. Specifically, we
asked participants questions about their opinions of the PSAs, as well as their perceptions of the
ad sponsorâs manipulative intent. Regarding the former, participants were asked to respond to the
question, âWhat is your opinion of the PSA you just viewed?,â followed by 78point item pairs
including: âNot at all informative/Very informative,â âNot at all persuasive/Very persuasive,â
âNot at all useful/Very useful,â âVery offensive/Not at all offensive,â and âUnfair/Fair.â For all
of these opinion measures, larger values represented more favorable appraisals of the PSA.
Perceptions of manipulative intent were then collected, using a 78point Likert (âStrongly
disagreeâ (1)/âStrongly agreeâ (7)) scale. The statement âPlease answer the following questions
based on your opinion of the PSA,â was followed by six items adapted from Campbell (1995).
Page 23 of 59 Journal of Consumer Affairs
26. F
o
r
R
e
v
i
e
w
O
n
l
y
23
These items included: (1) âThe way this PSA tries to persuade people seems acceptable to me,â
(2) âThe sponsor of this PSA tries to manipulate me in ways that I donât like,â (3) âI was
annoyed by this PSA because the sponsor seemed to be trying to inappropriately manage or
control me,â (4) âI donât mind this PSA; the sponsor tried to be persuasive without being
excessively manipulative,â (5) âThis PSA was fair in what was said and shown,â and (6) âI think
that this PSA is unfair.â After reverse coding items 1, 4, and 5, a manipulative intent scale was
created (α = .84), with larger values representing greater (i.e., more negative) perceptions of
manipulative intent.
Results
! "
Results of a univariate analysis indicated that the PSAs containing mortality salience
primes were effective in priming thoughts of death, as intended (#(1, 220) = 3.57, < .05).
Follow8up tests showed that as compared to the control PSA condition ( = 3.91), the verbal (
= 4.52, (110) = 1.69, < .05), visual ( = 4.77, (115) = 2.52, < .01), and verbal/visual ( =
4.93, (113) = 2.89, < .01) PSA treatment conditions evoked significantly greater thoughts of
death.
#
A 2 (Verbal Cues: Present vs. Absent) x 2 (Visual Cues: Present vs. Absent) ANOVA on
the scale revealed no significant main effect for verbal cues (#(1,
220) = 1.83, = .18), visual cues (#(1, 220) = 1.76, = .19), or for their interaction (#(1, 220) =
1.22, = .27). Our expectation was that if statistically significant differences were to emerge,
Page 24 of 59
Journal of Consumer Affairs
27. F
o
r
R
e
v
i
e
w
O
n
l
y
24
they would most likely appear in comparisons between the control PSA and the primed PSAs.
Consistent with this expectation, results of a univariate ANOVA comparing the control (i.e., no
prime) condition to the three primed conditions (combined) revealed a significant main effect for
(#(1, 222) = 5.05, < .05), with the primed conditions generating
less favorable attitudes. We then used follow8up planned comparisons to assess attitudinal
responses to PSAs containing (1) both visual and verbal primes, (2) a visual prime only, (3) a
verbal prime only, and (4) no primes (the superscripts corresponding to the means in Table 2 are
relevant to these comparisons). As expected, results for showed a
statistically significant difference between each of the three mortality8salience8primed PSAs and
the control condition. Specifically, as compared to the control PSA condition ( AB = 1.75), the
verbal ( AB = 1.47), visual ( AB = 1.47), and verbal/visual ( AB = 1.44) PSA treatment
conditions generated less favorable attitudes toward texting while driving (all 8values less than
.05). There were no significant differences across the three mortality salience prime conditions
(all 8values greater than .05), as shown in Table 2.
[Insert Table 2 about here]
A 2 (Verbal Cues: Present vs. Absent) x 2 (Visual Cues: Present vs. Absent) ANOVA on
the scale revealed a significant main effect for visual cues (#(1, 220) =
5.82, < .05), but no significant main effect for verbal cues (#(1, 220) = .60, = .44), or for the
interaction of verbal and visual cues (#(1, 220) = 1.47, = .23). Similar to the analyses reported
for AB above, results of a univariate ANOVA comparing the control (i.e., no prime) condition to
the three primed conditions revealed a significant main effect for (#(1,
222) = 6.90, < .01), with the primed condition yielding reduced intentions to text and drive.
Planned comparisons testing the mean differences between each of the four conditions were
Page 25 of 59 Journal of Consumer Affairs
28. F
o
r
R
e
v
i
e
w
O
n
l
y
25
subsequently conducted to determine the effects of the mortality salience primes on participantsâ
texting8while8driving behavioral intentions. Results showed a significant difference between the
control PSA and two of the three mortality salience8primed PSAs, offering some support for our
expectations. The visual ( BI = 1.63) and verbal/visual ( BI = 1.73) PSA treatments generated
significantly lower intentions to text while driving (both 8values less than .05) than the control
condition ( BI = 2.43). However, no statistically significant difference emerged between the
verbal mortality salience prime PSA condition ( BI = 1.99) and the control condition, though the
means were in the expected direction (see Table 2).
Similar to statistical procedures followed for the AB and BI measures, a univariate
ANOVA was conducted on the manipulative intent measure by comparing the control condition
to the three primed PSA conditions. To be deemed effective, we would expect that participantsâ
perceptions of the PSAs would be generally favorable (i.e., that participants would not perceive
the PSAs as being manipulative), and that manipulative intent would not vary significantly
across the four treatment conditions. Results of the univariate ANOVA revealed no significant
differences, in support of this notion (#(1, 222) = 1.31, = .25). Similarly, planned follow8up
comparisons between each of the four conditions indicated that perceptions of the manipulative
intent of the PSAs were no greater in any of the primed treatment conditions as compared to the
control (all 8values > .10). Additionally, these values were well below the scale midpoint of 4.0,
further suggesting that participants were not reacting in a negative/defensive way toward any of
the PSAsâ intended messages. Table 2 presents descriptive statistics and the results of these
analyses.
The results of our analysis of additional measures of participantsâ opinions of the PSAs
offer further support that the use of mortality salience primes did not adversely influence
Page 26 of 59
Journal of Consumer Affairs
29. F
o
r
R
e
v
i
e
w
O
n
l
y
26
opinions. Specifically, the PSAs featuring verbal and/or visual primes were perceived to be no
more offensive or unfair ( 8values > .10) than the control PSA. All values were well above the
scale midpoint, documenting participantsâ general positivity toward the PSAs. The positive
appraisal of the primed PSAs (as compared to the control) was also observed for the opinion
measures of persuasiveness, informativeness and usefulness; the primed PSAs were perceived to
be more persuasive, informative, and useful than the control PSA (all 8values < .05).
GENERAL DISCUSSION
Key Findings
Our initial exploratory study (Study 1) sought to determine some of the underlying
reasons why many young drivers tend to routinely text and drive â helping us to gain several
important insights about driversâ perceptions and behaviors in regards to this growing social
problem. Importantly, respondents who admitted that they often text while driving tended to
rationalize their actions by arguing that they were able to drive safely â even though they were
aware of the inherent dangers and potential consequences of their behavior. Moreover,
death/dying was a common theme found in many participantsâ responses. In Study 2, we
followed up on these observations by examining the impact of mortality salience primes within
public service announcements. Consistent with our theoretical expectations, results showed that
participants who were exposed to PSAs featuring verbal and/or visual mortality priming cues
generally held less favorable attitudes and reported reduced intentions to text while driving in the
future.
Page 27 of 59 Journal of Consumer Affairs
30. F
o
r
R
e
v
i
e
w
O
n
l
y
27
Theoretical and Practical Implications
The general findings of our studies are consistent with the results of meta8analyses that
have examined the persuasive impact of fear8arousing communications on perceptions of health
risks (de Hoog, Stroebe, and de Wit 2007; Witte and Allen 2000). In particular, our Study 2
results, much like those reported in recent meta8analyses, indicate that both vivid images and
written severity information may effectively alter respondentsâ attitudes and intentions,
consistent with an advertiserâs advocated position. However, given the relative advantage
symbols often enjoy over abstract or even concrete words, one might argue that the use of visual
mortality salience primes may be a preferred strategy for social marketers to consider. Symbols
are often universally understood and leave little room for alternative interpretations (Krampen
1965), making them a potentially powerful means of evoking a desired feeling or emotion.
Furthermore, the use of death symbols in PSAs is easy and inexpensive to implement, thereby
increasing their applicability by social marketers. Our findings regarding post8exposure
behavioral intentions offer some support for the use of symbols in this regard. After being
exposed to PSAs containing a skull8and8crossbones symbol (with or without an accompanying
written message), study participants reported significantly lower intentions to text and drive, as
compared to participants who were exposed to the PSA featuring only a verbal mortality salience
priming cue.
The results of Study 2 also support recent reports in the fear appeals literature that have
challenged the notion of an inverted8U relationship between levels of fear and persuasion. Our
findings, similar to those noted by others (e.g., de Hoog, Stroebe, and de Wit 2007; Sutton 1982;
Witte and Allen 2000), suggest that the use of promotional campaigns featuring relatively strong
emotional references to death/dying may be an effective persuasive technique. In Study 2, the
Page 28 of 59
Journal of Consumer Affairs
31. F
o
r
R
e
v
i
e
w
O
n
l
y
28
visual and/or verbal cues in our PSAs prompted relatively high levels of mortality salience, and,
nonetheless, were shown to effectively influence attitudes and intentions. Although our study did
not specifically test low, moderate, and high levels of fear, our findings are consistent with the
results of a research synthesis by Sutton (1982) that found high levels of fear are positively
associated with changes in intentions and behavior.
In addition, while our manipulation8checks document that our mortality8salience8primed
PSAs (as compared to the control condition) generated significantly greater thoughts of death,
we observed that perceptions of the sponsorâs manipulative intent did not differ significantly
across conditions. This suggests that psychological reactance did not adversely impact
participantsâ responses to the primed PSAs â a very real concern advertisers often face when
trying to reach their audiences, without appearing to be overly forceful and manipulative in their
persuasive attempts (Campbell 1995). These findings were further supported with additional
measures of participantsâ perceptions of the fairness and inoffensiveness of the PSAs, which
indicated that references to death (in either visual or verbal form) did not appear to unduly skew
recipientsâ perceptions of the PSA in a negative manner.
While one might argue that Study 2âs findings are based upon fear8evoking stimuli that
were compared to an experimenter8developed control condition that may not be a meaningful
reference point from which to compare the effects of the mortality8salience8primed PSAs, our
Study 1 data help to refute such a claim. Specifically, participants in Study 1 were not exposed to
any PSAs; they were simply asked to respond to questions regarding their attitudes and
intentions to text while driving in the future, along with other follow8up measures. Nevertheless,
these participantsâ mean responses to the attitude and behavioral intention measures were
strikingly similar to the responses obtained from Study 2 participants who had been exposed to
Page 29 of 59 Journal of Consumer Affairs
32. F
o
r
R
e
v
i
e
w
O
n
l
y
29
the control PSA.1
This pattern of results helps support the notion that the control PSA in Study 2
was a meaningful baseline from which to compare the additive effect of incorporating mortality
salience priming cues.
Although our investigation is preliminary in nature and more research is sorely needed on
this topic, our findings offer some hope for social marketers considering the effective use of
promotional campaigns to help curb the rising tide of texting while driving. With the nationâs
growing concern over this epidemic, sources of funding and support have become more readily
available. For example, the U.S. Department of Transportationâs National Highway Traffic
Safety Administration (NHTSA) recently announced a grant program that earmarks $17.5
million in FY 2013 funding to encourage states to educate the public via media campaigns and to
help enforce texting8while8driving laws (U.S. Department of Transportation 2012). Likewise,
companies such as telecommunications giant AT&T have pledged financial resources to support
anti8texting campaigns throughout the country (Crouch 2013). Moreover, the Ad Council, along
with State Attorneys General Offices and the NHTSA, have developed and continue to fund
multi8media (TV, radio, outdoor, and web8based) PSAs that target young adult drivers who
comprise a significant segment of distracted8driving offenders (NHTSA 2013). Our results
suggest that the use of fear appeals featuring mortality salience primes may be a worthwhile
avenue for state and federal agencies, as well as private enterprises, to pursue.
Limitations and Future Research Directions
As previously noted, our findings suggest that embedding verbal and/or visual references
to death in PSAs may effectively reduce participantsâ intentions to text and drive in the future.
1. Mean responses to the 78point AB and BI measures in Study 1 were 1.77 and 2.59,
respectively; responses to the control PSA in Study 2 were 1.75 and 2.43, respectively.
Page 30 of 59
Journal of Consumer Affairs
33. F
o
r
R
e
v
i
e
w
O
n
l
y
30
However, texting while driving has proven to be a challenging behavior to curb, as it continues
to plague our nationâs roads. Hence, we acknowledge that while Study 2 participants who were
exposed to the mortality8salience8primed PSAs reported reduced intentions to text and drive,
these intentions might not manifest as actual reductions in texting and driving behavior. In
another health8related context, Kees et al. (2006, 221) similarly recognized that while smokers
might indicate in an experimental setting that they intend to quit smoking, âthese effects may not
generalize to long8term smoking8related behaviorâ as stated intentions to quit âmight be different
from actively trying to quit or actually quitting.â
Relatedly, extant research suggests that texting while driving may be addictive for some
drivers (Bianchi and Philips 2005; Takao, Takahashi, and Kitamura 2009), thus leading to
additional challenges for social marketers trying to impact behaviors through persuasive
communications. Drivers who are addicted to texting and driving would likely perceive low self8
efficacy when exposed to a PSA designed to reduce texting8while8driving behaviors. The
literature and our conceptual model suggest that under such circumstances, fear appeals would
prove to be ineffectual, as addicted drivers would probably respond by following a control
(as opposed to a control) strategy, whereby they would likely use denial and/or
rationalization to reduce their fear (Witte 1992; Witte et al. 2001). Future research should further
explore the addictive nature of texting while driving, and focus on developing a more
comprehensive psychological profile of addicted drivers.
Future researchers might also consider exploring whether other types of fear appeals (i.e.,
employing manipulations other than mortality salience primes) might be effective in altering
recipientsâ texting8while8driving behaviors. One such possibility is to explore whether the threat
of âsocial exclusionâ (Charry and Demoulin 2012) may reduce the incidence of texting while
Page 31 of 59 Journal of Consumer Affairs
34. F
o
r
R
e
v
i
e
w
O
n
l
y
31
driving. The fear of being ostracized by their peers may prove to be a significant deterrent among
drivers who text and drive. Our Study 1 findings show that drivers who do not engage in this
risky behavior tend to feel very strongly against it â suggesting they may be effective role
models in promoting a change of behavior among peer group members.
Furthermore, we recognize that in our second study where manipulations were employed,
messages were negatively8framed (i.e., they emphasized the negative consequences of texting
while driving), and were presented in the form of print PSAs. Future researchers should attempt
to replicate and extend our findings using different stimuli and procedures. For example,
extensions to the present research could explore the effectiveness of positively8framed messages
on individualsâ texting8and8driving attitudes and behaviors (see, e.g., Algie 2011, regarding road
safety advertising appeals, and Hastings, Stead, and Webb 2004, regarding positive
reinforcement appeals). Moreover, extensions to the current investigation may wish to explore
the effectiveness of death symbols in PSAs appearing in other media. In particular, video PSAs
appearing on television, on various social media sites (such as Facebook and Twitter), and
disseminated through mobile video services (such as Vine) which are frequently visited by
young adults, would appear to be useful avenues worth pursuing.
In addition, social marketers should consider the context within which targeted
individuals will be exposed to messages aimed to deter them from texting while driving. Wong
and Householder (2008) found that individualsâ moods, as induced by the type of TV
programming they were viewing, impacted how viewers processed the persuasive message of an
anti8smoking PSA. Specifically, they found that when respondents were in a positive mood, they
were more likely to process peripheral message cues, as opposed to the arguments presented in
the message. In contrast, when respondents were in a negative mood, they processed the
Page 32 of 59
Journal of Consumer Affairs
35. F
o
r
R
e
v
i
e
w
O
n
l
y
32
arguments more systematically. Therefore, social marketers should carefully consider the context
and/or nature of the program in which their PSAs will be placed.
In conclusion, although it is a timely and relevant topic, consumer8focused journals to
date have published very little research related to texting while driving (see Lennon, Rentfro, and
OâLeary 2010 for an exception). We believe this is an oversight that should be addressed, as
social marketers may be best equipped to apply academic findings to more effectively deter
individuals from texting while driving. It is our hope that our studies will provide social
marketers useful insights to combat this critical issue facing our nation today, and will serve as a
catalyst for additional, much8needed research on this topic.
Page 33 of 59 Journal of Consumer Affairs
36. For Review
Only
33
APPENDIX
Study 2 Stimuli
Control condition Verbal prime condition
Visual prime condition Verbal and visual prime condition
Page 34 of 59
Journal of Consumer Affairs
37. F
o
r
R
e
v
i
e
w
O
n
l
y
34
REFERENCES
Ajzen, Icek. 1991. The Theory of Planned Behavior. $ % &
, 50 (2): 1798211.
Ajzen, Icek, and Martin Fishbein. 1980. '
. Englewood Cliffs, NJ: Prentice8Hall.
Algie, Jennifer. 2011. The Importance of Fear Reduction in Fear8Based Road Safety Advertising
Appeals. ! ( 22 (4): 998105.
Atchley, Paul, Stephanie Atwood, and Aaron Boulton. 2011. The Choice to Text and Drive in
Younger Drivers: Behavior May Shape Attitude. ) , 43 (1):
1348142.
Bianchi, Adriana, and James G. Phillips. 2005. Psychological Predictors of Problem Mobile
Phone Use. ! * ) , 8 (1): 39851.
Brehm, Jack W. 1966. + , New York: Academic Press.
Brehm, Jack W., and Sharon Stephens Brehm. 1981. - +
# ! . San Diego, CA: Academic Press.
Burke, Brian L., Andy Martens, and Erik H. Faucher. 2010. Two Decades of Terror Management
Theory: A Meta8Analysis of Mortality Salience Research.
, 14 (2): 1558195.
Campbell, Margaret C. 1995. When Attention8Getting Advertising Tactics Elicit Consumer
Inferences of Manipulative Intent: The Importance of Balancing Benefit and Investments.
! , 4 (3): 2258254.
Campbell, W. Keith, Adam S. Goodie, and Joshua D. Foster. 2004. Narcissism, Confidence, and
Risk Attitude. " , 17 (4): 2978311.
Page 35 of 59 Journal of Consumer Affairs
38. F
o
r
R
e
v
i
e
w
O
n
l
y
35
Charry, Karine M. and Nathalie T.M. Demoulin. 2012. Behavioural Evidence for the
Effectiveness of Threat Appeals in the Promotion of Healthy Food to Children.
, 31 (4): 773â94.
Consumer Reports. 2012. Distracted Driving Puts Young Drivers at Risk.
http://www.consumerreports.org/cro/magazine/2012/06/distracted8driving8puts8young8
drivers8at8risk/index.htm.
Cotte, June, Robin A. Coulter, and Melissa Moore. 2005. Enhancing or Disrupting Guilt: The
Role of Ad Credibility and Perceived Manipulative Intent. ,
58 (3): 3618368.
Crouch Ian. 2013. Why AT&T is Talking About Texting and Driving. + . " ,
http://www.newyorker.com/online/blogs/currency/2013/10/why8att8is8talking8about8
texting8and8driving.html.
Dahl, Darren W., Kristina D. Frankenberger, and Rajesh V. Manchanda. 2003. Does It Pay to
Shock? Reactions to Shocking and Nonshocking Advertising Content among University
Students. , 43 (3): 2688280.
Dahl, Ronald E. 2008. Biological, Developmental, and Neurobehavioral Factors Relevant to
Adolescent Driving Risks. , 35 (3S): S2788S284.
de Hoog, Natascha, Wolfgang Stroebe, and John B. F. de Wit. 2007. The Impact of Vulnerability
to and Severity of a Health Risk on Processing and Acceptance of Fear8Arousing
Communications: A Meta8Analysis. / , 11 (3): 258885.
Dingus, Thomas A., Richard J. Hanowski, and Sheila G. Klauer. 2011. Estimating Crash Risk.
- + 0 & # , 19 (8): 8812.
Page 36 of 59
Journal of Consumer Affairs
39. F
o
r
R
e
v
i
e
w
O
n
l
y
36
Drews, Frank A., Hina Yazdani, Celeste N. Godfrey, Joel M. Cooper, and David L. Strayer.
2009. Text Messaging During Simulated Driving. & # - +
& # , 51 (5): 7628770.
Elliott, Barry. 1993. Road Safety Mass Media Campaigns: A Meta Analysis. Research Report
(CR118), Department of Transport and Communications.
Flora, June A. and Edward W. Maibach. 1990. Cognitive Responses to AIDS Information: The
Effects of Issue Involvement and Message Appeal. ! , 17 (6): 7598
774.
Greenberg, Jeff, Tom Pyszczynski, and Sheldon Solomon. 1986. The Causes and Consequences
of the Need for Self8Esteem: A Terror Management Theory. In *
, edited by Roy F. Baumeister (1898212). New York: Springer8Verlag.
Greenberg, Jeff, Tom Pyszczynski, and Sheldon Solomon, Abram Rosenblatt, Mitchell Veeder,
Shari Kirkland, and Deborah Lyon. 1990. Evidence for Terror Management Theory II: The
Effects of Mortality Salience on Reactions to Those Who Threaten or Bolster the Cultural
Worldview. , 58 (2): 3088318.
Harrison, Marissa A. 2011. College Students' Prevalence and Perceptions of Text Messaging
While Driving. , 43: 151681520.
Hastings, Gerard, Martine Stead, and John Webb. 2004. Fear Appeals in Social Marketing
Strategic and Ethical Reasons for Concern. ) " , 21 (11): 961886.
Highway Loss Data Institute. 2010. Texting Laws and Collision Claim Frequencies. &
1 , 27 (11). Arlington, VA: Highway Loss Data Institute.
Page 37 of 59 Journal of Consumer Affairs
40. F
o
r
R
e
v
i
e
w
O
n
l
y
37
Hosking, Simon, Kristie Young, and Michael Regan. 2006. The Effects of Text Messaging on
Young Novice Driver Performance. Monash University Accident Research Centre: The
National Roads and Motorists' Association Motoring and Services.
Hovland, Carl I., Irving L. Janis, and Harold H. Kelly. 1953. ! .
New Haven, CT: Yale University Press.
Just, Marcel Adam, Timothy A. Keller, and Jacquelyn Cynkar. 2008. A Decrease in Brain
Activation Associated with Driving When Listening to Someone Speak. ,
1205: 70880.
Kees, Jeremy, Scot Burton, J. Craig Andrews, and John Kozup. 2006. Tests of Graphic Visuals
and Cigarette Package Warning Combinations: Implications for the Framework Convention
on Tobacco Control. * ) " , 25 (2): 2128223.
Kees, Jeremy, Scot Burton, J. Craig Andrews, and John Kozup. 2010. Understanding How
Graphic Pictorial Warnings Work on Cigarette Packaging. * )
" , 29 (2): 2658276.
Keller, Punam Anand. 1999. Converting the Unconverted: The Effect of Inclination and
Opportunity to Discount Health8Related Fear Appeals. , 84
(3): 403815.
Keller, Punam Anand, and Donald R. Lehmann. 2008. Designing Effective Health
Communications: A Meta8Analysis. * ) " , 27 (2): 117830.
Krampen, Martin. 1965. Signs and Symbols in Graphic Communication. 0 , 62:
1831.
Page 38 of 59
Journal of Consumer Affairs
41. F
o
r
R
e
v
i
e
w
O
n
l
y
38
Krisher, Howard P., Susan A. Darley, and John M. Darley. 1973. Fear8Provoking
Recommendations, Intentions to Take Preventive Actions, and Actual Preventive Actions.
, 26 (2): 301808.
Lee, Suzanne, Sheila Klauer, Erik Olsen, Bruce Simons8Morton, Thomas Dingus, David
Ramsey, and Marie Ouimet. 2008. Detection of Road Hazards by Novice Teen and
Experienced Adult Drivers. + , 2078: 26832.
Lennon, Ron, and Randall Rentfro. 2010. Are Young Adults Fear Appeal Effectiveness Ratings
Explained by Fear Arousal, Perceived Threat, and Perceived Efficacy?
" , 6 (1): 58865.
Lennon, Ron, Randall Rentfro, and Bay O'Leary. 2010. Social Marketing and Distracted Driving
Behaviors among Young Adults. " , 14 (2): 958113.
Leventhal, Howard. 1970. Findings and Theory in the Study of Fear Communications. In
2 , Vol. 5., edited by Berkowitz Leonard (1198
187). New York: Academic Press.
Lewis, I. M., B. Watson, and K. M. White. 2010. Response Efficacy: The Key to Minimizing
Rejection and Maximizing Acceptance of Emotion8Based Anti8Speeding Messages.
) , 42 (2): 4598467.
Magid, Larry. 2009. Shocking Stats on Texting While Driving. CBS News.
http://www.cbsnews.com/stories/2009/08/29/scitech/pcanswer/main5274193.shtml.
National Highway Traffic Safety Administration. 2013. Official US Government Website for
Distracted Driving. http://www.distraction.gov.
National Safety Council. 2010. National Safety Council Estimates that At Least 1.6 Million
Crashes Each Year Involve Drivers Using Cell Phones and Texting.
Page 39 of 59 Journal of Consumer Affairs
42. F
o
r
R
e
v
i
e
w
O
n
l
y
39
http://www.nsc.org/pages/nscestimates16millioncrashescausedbydriversusingcellphonesan
dtexting.aspx.
Nelson, Erik, Paul Atchley, and Todd D. Little. 2009. The Effects of Perception of Risk and
Importance of Answering and Initiating a Cellular Phone Call While Driving.
) , 41 (3): 4388444.
Nemme, Heidi E., and Katherine M. White. 2010. Texting While Driving: Psychosocial
Influences on Young People's Texting Intentions and Behaviour. )
, 42 (4): 125781265.
Obermiller, Carl, Eric Spangenberg, and Douglas L. MacLachlan. 2005. Ad Skepticism: The
Consequences of Disbelief. , 34 (3): 7817.
Occupational Safety & Health Administration. 2010. Distracted Driving.
http://www.osha.gov/distracted8driving/index.html.
Passyn, Kirsten, and Mita Sujan. 2006. Self8Accountability Emotions and Fear Appeals:
Motivating Behavior. ! , 32 (4): 5838589.
Paulhus, Delroy L., P. D. Harms, M. Nadine Bruce, and Daria C. Lysy. 2003. The Over8
Claiming Technique: Measuring Self8Enhancement Independent of Ability.
, 84 (4): 8908904.
Pavia, Teresa M., and Marlys J. Mason. 2004. The Reflexive Relationship between Consumer
Behavior and Adaptive Coping. ! , 31 (2): 4418454.
Perse, Elizabeth M., Amy I. Nathanson, and Douglas M. McLeod. 1996. Effects of Spokesperson
Sex, Public Service Announcement Appeal, and Involvement on Evaluations of Safe8Sex
PSAs. & ! , 8 (2): 1718189.
Page 40 of 59
Journal of Consumer Affairs
43. F
o
r
R
e
v
i
e
w
O
n
l
y
40
Petty, Richard E. and John T. Cacioppo. 1986. ! - !
! . New York: Springer8Verlag.
Preacher, Kristopher J. and Andrew F. Hayes. 2004. SPSS and SAS Procedures for Estimating
Indirect Effects in Simple Mediation Models. ( (
! , 36 (4): 717â31.
Preacher, Kristopher J., and Andrew F. Hayes. 2008. Asymptotic and Resampling Strategies
for Assessing and Comparing Indirect Effects in Multiple Mediator Models.
, 40 (3): 8798891.
Quinn, Valerie, Tony Meenaghan, and Teresa Brannick. 1992. Fear Appeals: Segmentation Is
the Way to Go. , 11, 3558366.
Ranney, Thomas A., G. H. Scott Baldwin, Ed Parmer, John Martin, and Elizabeth N. Mazzae.
2012. Distraction Effects of In8Vehicle Tasks Requiring Number and Text Entry Using
Auto Allianceâs Principle 2.1B Verification Procedure, Report No. DOT HS 811 571.
National Highway Traffic Safety Administration.
Reardon, James, Chip Miller, Bram Foubert, Irena Vida, and Liza Rybina. 2006. Antismoking
Messages for the International Teenage Segment: The Effectiveness of Message Valence
and Intensity across Different Cultures. " , 14 (3): 1158
138.
Redelmeier, Donald A., and Robert J. Tibshirani. 1997. Association Between Cellular8Telephone
Calls and Motor Vehicle Collisions. , 336 (7): 4538458.
Rogers, Ronald W. 1975. A Protection Motivation Theory of Fear Appeals and Attitude Change.
+ y, 91 (1): 938114.
Page 41 of 59 Journal of Consumer Affairs
44. F
o
r
R
e
v
i
e
w
O
n
l
y
41
Rogers, Ronald W. 1983. Cognitive and Physiological Processes in Fear Appeals and Attitude
Change: A Revised Theory of Protection Motivation. In -
* ", edited by John T. Cacioppo and Richard E. Petty (1538176). New York:
Guilford Press.
Routledge, Clay, and Jacob Juhl. 2010. When Death Thoughts Lead to Death Fears: Mortality
Salience Increases Death Anxiety for Individuals Who Lack Meaning in Life. !
, 24 (5): 8488854.
Ruiter, Robert A. C., Charles Abraham, and Gerjo Kok. 2001. Scary Warnings and Rational
Precautions: A Review of the Psychology of Fear Appeals. ) & , 16 (6):
6138630.
Shehryar, Omar, and David M. Hunt. 2005. A Terror Management Perspective on the
Persuasiveness of Fear Appeals. ! , 15 (4): 2758287.
Spears, Nancy, and Surendra N. Singh. 2004. Measuring Attitude toward the Brand and Purchase
Intentions. ! ) , 26 (2): 53866.
Steelman, Zackary, Amr Soror, Moez Limayem, and Dan Worrell. 2012. Obsessive Compulsive
Tendencies as Predictors of Dangerous Mobile Phone Usage. In
! , Paper 9.
Sternthal, Brian and C. Samuel Craig. 1974. Fear Appeals: Revisited and Revised.
! , 1 (3): 22834.
Strayer, David L., Frank A. Drews, and Dennis J. Crouch. 2006. A Comparison of the Cell
Phone Driver and the Drunk Driver. & # , 48 (2): 3818391.
Page 42 of 59
Journal of Consumer Affairs
45. F
o
r
R
e
v
i
e
w
O
n
l
y
42
Sutton, Stephen R. 1982. Fear8Arousing Communications: A Critical Examination of Theory and
Research. In , edited by J. Richard Eiser,
Chichester (3038307). UK: John Wiley & Sons Ltd.
Takao, Motoharu, Susumu Takahashi, and Masayoshi Kitamura. 2009. Addictive Personality and
Problematic Mobile Phone Use. ! * ) , 12 (5): 501807.
Tanner, John F., Jr., James B. Hunt, and David R. Eppright. 1991. The Protection Motivation
Model: A Normative Model of Fear Appeals. " , 55 (3): 36845.
Taylor, Charles R., Thomas A. Claus, and Susan L. Claus. 2005. $ 3
" . Washington, DC: U.S. Small Business Administration.
Terblanche8Smit, Marlize and Nic S. Terblanche. 2010. Race and Attitude Formation in
HIV/AIDS Fear Advertising. , 63 (2): 1218126.
Tien, Cheryl, and Ian Phau. 2010. Conceptualizing the Mediating Role of Inferences of
Manipulative Intent Between Consumer Skepticism and Product Judgment. In
! ( edited by Harry Timmermans,
Istanbul: Recent Advances in Retailing and Services Science.
Tison, Julie, Neil Chaudhary, and Lisa Cosgrove. 2011. National Phone Survey on Distracted
Driving Attitudes and Behaviors. Research Report, Department of Transportation, National
Highway Traffic Safety Association.
Twenge, Jean M., Sara Konrath, Joshua D. Foster, W. Keith Campbell, and Brad J. Bushman.
2008. Egos Inflating Over Time: A Cross8Temporal Meta8Analysis of the Narcissistic
Personality Inventory. , 76 (4): 8758902.
U.S. Department of Transportation. 2012. http://www.distraction.gov/content/press8
release/2012/08822.html.
Page 43 of 59 Journal of Consumer Affairs
46. F
o
r
R
e
v
i
e
w
O
n
l
y
43
U.S. Small Business Administration. 2003. * "- & * ".
Washington, DC: U.S. Small Business Administration.
Wickman, Mary E., Nancy Lois Ruth Anderson, and Cindy Smith Greenberg. 2008. The
Adolescent Perception of Invincibility and its Influence on Teen Acceptance of Health
Promotion Strategies. , 23 (6): 4608468.
Witte, Kim. 1992. Putting the Fear Back into Fear Appeals: The Extended Parallel Process
Model. ! , 59 (4): 3298349.
Witte, Kim, and Mike Allen. 2000. A Meta8Analysis of Fear Appeals: Implications for Effective
Public Health Campaigns. & ) , 27 (5): 5918615.
Witte, Kim, Gary Meyer, and Dennis Martell. 2001. & " - 3* 3
/ . Thousand Oaks, CA: Sage.
Wong, Norma C.H., and Brian Householder. 2008. Mood and Ad Processing: Examining the
Impact of Program8Induced Moods on Subsequent Processing of an Antismoking Public
Service Announcement. ! , 59 (4): 4028414.
Zhao,Xinshu, John G. Lynch Jr., and Qimei Chen. 2010. Reconsidering Baron and Kenny:
Myths and Truths about Mediation Analysis. ! , 37 (2): 1978
206.
Page 44 of 59
Journal of Consumer Affairs
47. For Review
Only
44
TABLE 1
4
Notes: a
Tabled are mean values with standard deviations in parentheses (âStrongly disagreeâ (1)/âStrongly agreeâ (7)). âYesâ (âNoâ) refers to
scores above (below) the median of 1.5 on the 78point scale.
b
The 8values correspond to tests of differences between the âYesâ and âNoâ groups.
c
Correlation coefficients between evaluative judgments and the continuous measure of .
d
Correlation coefficients between evaluative judgments and the continuous measure of .
*** < .001.
Plan to Text8and8Drive over Next Month?a
Correlations
Evaluative Judgments of Hypothetical Situations Yes No 8valueb
BIc
ABd
There are times when it is OK to text while driving 3.87 (1.94) 1.86 (1.46) 11.03*** .63*** .61***
It is OK for me to text while driving when stopped at a red light 5.31 (1.62) 2.88 (1.85) 13.21*** .60*** .50***
It is OK for me to text while driving if there is an emergency 4.57 (1.99) 2.37 (1.84) 10.88*** .57*** .53***
It is OK for me to text while driving if there is not a lot of traffic on the road 3.21 (1.85) 1.41 (.97) 11.46*** .64*** .67***
I would text while driving to get/give directions 3.59 (1.91) 1.60 (1.28) 11.52*** .64*** .56***
I would text while driving if I wanted to know something instantly 3.40 (2.02) 1.26 (.65) 13.35*** .77*** .60***
I would text while driving if I thought I could get away with it 3.56 (1.97) 1.25 (.76) 14.52*** .76*** .63***
I would text while driving if I was responding to a message from my family members 3.33 (1.86) 1.23 (.62) 14.14*** .78*** .65***
I would text while driving if I was responding to a message from my friends 3.21 (1.86) 1.20 (.58) 13.62*** .78*** .65***
I would NOT text while driving under any circumstances 3.52 (1.94) 6.15 (1.48) 814.39*** 8.71*** 8.51***
I would NOT text while driving even if I was bored 4.37 (1.99) 6.49 (1.36) 811.73*** 8.62*** 8.48***
If I text while driving I will kill myself in a crash 4.52 (1.62) 5.64 (1.35) 87.12*** 8.40*** 8.47***
If I text while driving I will kill others in a crash 4.51 (1.69) 5.68 (1.37) 87.13*** 8.40*** 8.45***
Page 45 of 59 Journal of Consumer Affairs
48. For Review
Only
45
TABLE 2
5 - ! 6 7 8 *
Notes: Tabled are mean values with standard deviations in parentheses; for each row, superscripts indicate statistically
significant differences ( < .05 or better) according to post hoc LSD comparisons. For example, the mean attitude
toward texting while driving for cell A (Verbal Absent/Visual Absent) is statistically different from the means for
cells B, C, and D.
Mortality Salience Primes
Measures
A: Verbal
Absent/Visual
Absent
B: Verbal
Present/Visual
Absent
C: Verbal
Absent/Visual
Present
D: Verbal
Present/Visual
Present
9 8 * -
Attitude Toward Texting While Driving 1.75 (.95)B,C,D
1.47 (.85)A
1.47 (.76)A
1.44 (.87)A
Behavioral Intentions 2.43 (1.96)C,D
1.99 (1.80) 1.63 (1.18)A
1.73 (1.49)A
-
Manipulative Intent 2.48 (.96) 2.23 (.87) 2.41 (.96) 2.29 (1.15)
The Public Service Announcement was:
Not offensive 5.82 (1.52) 6.21 (1.19)C,D
5.54 (1.68)B
5.64 (1.74)B
Fair 5.38 (1.22)B
5.98 (1.21)A
5.75 (1.29) 5.67 (1.60)
Persuasive 3.43 (1.93)B,C,D
4.62 (1.96)A
4.14 (1.88)A,D
5.11 (1.83)A,C
Informative 3.33 (1.93)B,C,D
5.02 (1.73)A,C
3.88 (1.68)A,B,D
5.27 (1.66)A,C
Useful 3.73 (1.97)B,C,D
5.10 (1.73)A,C
4.32 (1.67)A,B,D
5.20 (1.80)A,C
Page 46 of 59
Journal of Consumer Affairs
49. For Review
Only
46
.70*** .76***
.05
FIGURE 1
+ + 2 : * 6 47
Notes: Model ( 2
= .64) reports standardized regression coefficients.
*** < .001.
Attitude
Toward Texting
While Driving
(AB)
Behavioral
Intentions
(BI)
Evaluative
Judgments
(EJ)
Page 47 of 59 Journal of Consumer Affairs
50. For Review
Only
47
FIGURE 2
! - 2 ! 2 + 2 3: 3
No Perceived Threat
(No Response)
Notes: Model adapted from Witte (1992, p. 338).
Mortality Salience
Prime (Verbal
and/or Visual)
Susceptibility
Severity
(Self8Efficacy,
Response
Efficacy)
!
!
(Maladaptive
Changes)
Attitudes &
Intentions to
Text and
Drive
Message is
Discounted
( !
in Attitudes
& Intentions
to Text and
Drive)
Reactance
(Manipulative
Intent)
Attitudes &
Intentions to
Text and
Drive
(Boomerang
Effect)
"
#
(Susceptibility,
Severity)
!
(Adaptive
Changes)
Page 48 of 59
Journal of Consumer Affairs